Vectorization of Multiple Small Matrix Problems

Abstract

Multiple independent matrix problems of very small size appear in a variety of different fields. In this work, we study the implementation of elementary linear algebra subroutines so as to best use vectorizing compilers and vector hardware for multiple small problem instances. We cheek the performance improvement over the single-instance optimized codes on different vector supercomputers. We also describe how to automate the transformation of a single-instance linear algebra solver into a multiple instance solver